微软本地文本生成音频

把指定目录下所有txt文件，用Windows系统自带语音引擎(SAPI)转换为wav音频。

import os
import win32com.client

# ====== 配置区 ======
TXT_DIR = r"G:\test"           # 你的txt目录
OUTPUT_DIR = r"G:\test"        # 输出音频目录
VOICE_NAME = "Microsoft Huihui Desktop"  # 可换成 powershell -Command "(New-Object -ComObject SAPI.SpVoice).GetVoices() | ForEach-Object { $_.GetDescription() }"
RATE = 0      # 语速（-10 ~ 10）
VOLUME = 100  # 音量（0 ~ 100）
# ===================

os.makedirs(OUTPUT_DIR, exist_ok=True)

speaker = win32com.client.Dispatch("SAPI.SpVoice")
speaker.Voice = [v for v in speaker.GetVoices() if VOICE_NAME in v.GetDescription()][0]
speaker.Rate = RATE
speaker.Volume = VOLUME

for file in os.listdir(TXT_DIR):
    if file.lower().endswith(".txt"):
        txt_path = os.path.join(TXT_DIR, file)
        wav_path = os.path.join(OUTPUT_DIR, os.path.splitext(file)[0] + ".wav")

        with open(txt_path, "r", encoding="utf-8") as f:
            text = f.read().strip()

        if not text:
            print(f"⚠ 跳过空文件: {file}")
            continue

        print(f"🎤 正在生成: {file} → {wav_path}")
        stream = win32com.client.Dispatch("SAPI.SpFileStream")
        stream.Open(wav_path, 3, False)
        speaker.AudioOutputStream = stream
        speaker.Speak(text)
        stream.Close()

print("✅ 全部转换完成！")